ABSTRACT
OBJECTIVES: The present study compared the performance of the Lumipulse G Sars-CoV-2 Ag kit with the TaqPath COVID-19 RT-PCR CE IVD kit. METHODS: The study was conducted on 4266 naso-oropharyngeal swabs. Samples were subjected to antigen RT-PCR tests for the detection of Sars-CoV-2 and related variants. Statistical analyses were conducted in R software. RESULTS: We found 503 positives (including 138 H69-V70 deletion carriers) and 3763 negatives by RT-PCR, whereas 538 positives and 3728 negatives were obtained by antigen testing. We achieved empirical and binormal AU-ROCs of 0.920 and 0.990, accuracy of 0.960, sensitivity of 0.866, specificity of 0.973, positive and negative predictive values of 0.810 and 0.980. We obtained a positive correlation between viral loads and antigen levels (R2 = 0.81), finding a complete concordance for high viral loads (log10 copies/mL > 5.4). Antigen levels > 222 pg/mL were found to be reliable in assigning positive samples (p < 0.01). Concerning variant carriers, antigen test detected them with the same accuracy as other positive samples. CONCLUSIONS: Molecular and antigen tests should be evaluated regarding the prevalence of the area. In case of low prevalence, antigen testing can be employed as a first-line screening for the timely identification of affected individuals with high viral load, also if carriers of Sars-CoV-2 variants.